Category Not Found

1252 prompts

Sort:

Reduce p95 latency on math word problems Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a math word problems prompt on Claude Haiku 4 against p95 latency without regressing safety.

Reduce accuracy on multi-hop QA Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a multi-hop QA prompt on GPT-4o against accuracy without regressing safety.

Reduce F1 score on research synthesis Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a research synthesis prompt on Gemini 2.5 Pro against F1 score without regressing safety.

Reduce F1 score on legal brief summarization Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a legal brief summarization prompt on GPT-4.1 against F1 score without regressing safety.

Reduce factuality on incident post-mortems Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a incident post-mortems prompt on Gemini 2.0 Flash against factuality without regressing safety.

Reduce refusal rate on data pipeline debugging Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a data pipeline debugging prompt on o1-mini against refusal rate without regressing safety.

Reduce toolcall precision on resume screening Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a resume screening prompt on Claude 3.7 Sonnet against toolcall precision without regressing safety.

Reduce toolcall precision on math word problems Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a math word problems prompt on o3-mini against toolcall precision without regressing safety.

Reduce format-compliance rate on multi-hop QA Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a multi-hop QA prompt on Llama 3.3 70B against format-compliance rate without regressing safety.

Reduce hallucination rate on legal brief summarization Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a legal brief summarization prompt on Llama 3.1 405B against hallucination rate without regressing safety.

Reduce user satisfaction (CSAT) on incident post-mortems Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a incident post-mortems prompt on Command R+ against user satisfaction (CSAT) without regressing safety.

Reduce inter-judge agreement on data pipeline debugging Prompt via Textual Gradient Descent

Use Textual Gradient Descent to optimize a data pipeline debugging prompt on Claude Haiku 4 against inter-judge agreement without regressing safety.

🟠Claude

26151